Histwords — Diachronic Semantic Change in LDS Religious Language

Course: LING 581 | BYU Author: Jacob B. Fisher

Fork of williamleif/histwords — modified to study semantic change in LDS/religious language across historical corpora.

Modifications from Original

This fork extends the original Histwords toolkit to analyze diachronic semantic change in Latter-day Saint (LDS) scripture and religious language. Specifically:

• Applied the historical word embedding pipeline to religious text corpora
• Tracked semantic drift of theologically significant vocabulary over time
• Added back_track_centroid.py, track_neighbor_centroid.py, and similarity_neighbors_tracking.py for tracing word meaning trajectories
• Used tsne_viz.py for 2D visualization of embedding shifts

Original Project

An eclectic collection of tools for analyzing historical language change using vector space semantics (SGNS word embeddings trained on Google N-grams and COHA).

Original author: William Hamilton (wleif@stanford.edu) Project website: http://nlp.stanford.edu/projects/histwords Paper: Diachronic Word Embeddings Reveal Statistical Laws of Semantic Change (ACL 2016)

Usage

# Install dependencies
pip install -r requirements.txt

# Run example
bash example.sh
python example.py

Technologies

• Python, NumPy, scikit-learn, t-SNE, SGNS word embeddings